Error starting indexing: Cannot convert argument to a ByteString because the character at index 7 has a value of 8226 which is greater than 255.

[PavelA85]

App Version

3.28.1

API Provider

VS Code Language Model API

Model Used

text-embedding-3-small

Roo Code Task Links (Optional)

No response

🔁 Steps to Reproduce

1. start npx --package copilot-api@latest --registry=https://registry.npmjs.org -- copilot-api start --port 8081
2. start docker run -d --name qdrant-instance -p 6333:6333 qdrant/qdrant && echo 'Qdrant started'
3. setup indexing:

		"codebaseIndexConfig": {
			"codebaseIndexEnabled": true,
			"codebaseIndexQdrantUrl": "http://localhost:6333",
			"codebaseIndexEmbedderProvider": "openai-compatible",
			"codebaseIndexEmbedderBaseUrl": "",
			"codebaseIndexEmbedderModelId": "text-embedding-3-small",
			"codebaseIndexEmbedderModelDimension": 1536,
			"codebaseIndexSearchMinScore": 0.4,
			"codebaseIndexSearchMaxResults": 50,
			"codebaseIndexOpenAiCompatibleBaseUrl": "http://localhost:8081"

4. I get error: Error starting indexing: Cannot convert argument to a ByteString because the character at index 7 has a value of 8226 which is greater than 255.

💥 Outcome Summary

Expected to get indexing started, but getting error Error - Cannot convert argument to a ByteString because the character at index 7 has a value of 8226 which is greater than 255.

Only thing helps, is to wipe C:\Users\MyUser\AppData\Roaming\Code - Insiders\ folder
I am using VSCode Insiders 1.105.0-insider (user setup) version

Version: 1.105.0-insider (user setup)
Commit: 2a2f830103fd0f9b46ea56b203bc28e2e22541e9
Date: 2025-09-12T05:03:11.987Z
Electron: 37.3.1
ElectronBuildId: 12342881
Chromium: 138.0.7204.235
Node.js: 22.18.0
V8: 13.8.258.31-electron.0
OS: Windows_NT x64 10.0.22631

📄 Relevant Logs or Errors (Optional)

roo-cline extension activated - {"publisher":"RooVeterinaryInc","name":"roo-cline","version":"3.28.1","outputChannel":"Roo-Code"}
[Default Commands Migration] Migration already completed, skipping
Skipping migration of cline_custom_modes.json to custom_modes.json: old file not found
Skipping migration of cline_mcp_settings.json to mcp_settings.json: old file not found
No custom_modes.json found, skipping YAML migration
[CodeIndexManager] Error during background CodeIndexManager configuration/indexing for c:\Projects\MyProjects: Cannot convert argument to a ByteString because the character at index 7 has a value of 8226 which is greater than 255.
CloudService not ready, deferring cloud profile sync
[auth] Using WebAuthService
[auth] changeState: initializing -> attempting-session
[AutoImport] No auto-import settings path specified, skipping auto-import
[auth] changeState: attempting-session -> active-session
[auth] User in personal account context - not setting organization info
Error starting indexing: Cannot convert argument to a ByteString because the character at index 7 has a value of 8226 which is greater than 255.

[roomote]

I'm exploring a potential fix for this ByteString encoding issue. The error appears to be related to Unicode character handling (character 8226 is a bullet point •) when interfacing with the OpenAI-compatible API. I'll investigate the codebase indexing implementation and open a PR to share progress.

[PavelA85]

Based on your commit I have reentered dummy API Key in indexing settings, and it started indexing.
e.g. It was ******** and I have entered 123, as my setup does not have API keys and it started to index.

Good job @roomote-agent 💘

[roomote]

Opened a PR with my attempt at addressing this issue: #7960

The fix sanitizes non-ASCII characters in API keys before using them in HTTP headers, which should resolve the ByteString conversion error. The implementation includes comprehensive tests and a warning message to help users identify when their API keys contain problematic characters.

Feedback welcome!

[roomote]

Thank you @PavelA85 for confirming the workaround works! 🎉

I'm glad re-entering the API key resolved the immediate issue for you. The root cause was indeed the masked API key containing non-ASCII characters (the bullet points • used for masking).

PR #7960 provides a permanent fix by sanitizing non-ASCII characters in API keys before they're used in HTTP headers. Once merged, this should prevent the error from occurring even with masked keys.

Thanks again for your feedback and for testing the solution! 💘

[PavelA85]

Ohhh, UI uses not * for API key placeholder, but • - 8226 char and it if API key is empty, it provides ••••••••••• as API Key and fails to create ByteString